COPYRIGHT AND PRODUCT DIFFERENTIATION
Christopher S. Yoo
Introduction..................................................................................................... 1
I. The Traditional Approach to the Economics of Copyright: Public
Goods .............................................................................................................. 4
A. The Second-Best Nature of the Public Goods Equilibrium.................. 5
B. Puzzles that the Public Goods Approach Cannot Resolve ................... 9
1. The Questionable Assumption of Constant Marginal Costs ........... 9
2. The Absence of Natural Monopoly Despite Constantly
Declining Average Costs ...................................................................... 9
3. The Persistence of Market Power Notwithstanding the
Prevalence of Substitutes and the Absence of Supracompetitive
Profits ................................................................................................. 10
4. The Inability to Resolve the Tradeoff Between Static and
Dynamic Efficiency............................................................................ 12
II. The Differentiated Products Approach to Copyright: Symmetric
Preference Models ........................................................................................ 13
A. The Theory of Monopolistic Competition.......................................... 13
B. The Descriptive Advantages of Applying Monopolistic
Competition to Copyright ........................................................................ 16
1. The Irrelevance of Nonrivalry ....................................................... 16
2. Product Differentiation as a Means for Averting Natural
Monopoly ........................................................................................... 17
3. The Coexistence of Market Power, Free Entry, and the
Absence of Supracompetitive Profits ................................................. 18
C. The Welfare Implications of Monopolistic Competition.................... 19
1. A New Perspective on Static Efficiency ....................................... 20
2. The Formalization of Dynamic Efficiency.................................... 21
a. Appropriability as a Determinant of Dynamic Efficiency ....... 21
b. The Indirect Relationship Between Appropriability and
Static Efficiency ............................................................................ 24
c. Demand Diversion as a Countervailing Consideration ............ 25
d. Striking the Optimal Balance ................................................... 28
III. The Differentiated Products Approach to Copyright: Asymmetric
Preference Models ........................................................................................ 29
A. Relaxing the Symmetric Preferences Assumption ............................. 30
B. Implications ........................................................................................ 32
IV.Application of Product Differentiation Theory to Specific
Copyright Issues............................................................................................ 33
A. Facilitation of Price Discrimination ................................................... 33
B. The Fair Use Doctrine ........................................................................ 34
C. The Scope of Copyright Protection .................................................... 35
Conclusion .................................................................................................... 36
[Rough draft: please do not cite or quote without permission]
COPYRIGHT AND PRODUCT DIFFERENTIATION
Christopher S. Yoo*
INTRODUCTION
To date, two central themes have dominated the economic analysis of
copyright law. The first theme frames copyright as a tradeoff between two
opposing considerations. On the one hand are the social benefits that accrue
from disseminating copyrighted works as widely as possible. On the other is
the need to provide authors with sufficient incentive to incur the fixed costs
investments needed to create those works in the first place.1 This tradeoff
between “access” and “incentives” can be stated more formally as a tradeoff
between static efficiency, which focuses on ensuring that goods are allocated
so as to maximize total surplus, and dynamic efficiency, which focuses on
ensuring that the optimal number of goods are created.
The second theme is that informational products exhibit qualities
associated with public goods.2 In its most extreme form, this approach posits
*
Associate Professor of Law, Vanderbilt University. Financial support from the Vanderbilt
Dean’s Fund is gratefully acknowledged.
1
For example, the Supreme Court described copyright as requiring “a difficult balance between
the interests of authors and inventors in the control and exploitation of their writings and discoveries on
the one hand, and society’s competing interest in the free flow of ideas, information, and commerce on the
other hand.” Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 429 (1984). William
Landes and Richard Posner similarly observed in their landmark article on the economics of copyright:
Copyright protection . . . trades off the costs of limiting access to a work against the benefits of
providing incentives to create the work in the first place. Striking the correct balance between
access and incentives is the central problem in copyright law.
William M. Landes & Richard A. Posner, An Economic Analysis of Copyright Law, 18 J. LEGAL STUDS.
325, 326 (1999). For other representative statements appearing in the literature, see Stephen Breyer, The
Uneasy Case for Copyright: A Study of Copyright in Books, Photocopies, and Computer Programs, 84
HARV. L. REV. 281, 281-82 (1970); Julie E. Cohen, Copyright and the Perfect Curve, 53 VAND. L. REV.
1799, 1800-01 (2000); William W. Fisher III, Reconstructing the Fair Use Doctrine, 101 HARV. L. REV.
1659, 1700-03 (1988); S.J. Liebowitz, Copyright Law, Photocopying and Price Discrimination, 8 RES. L.
& ECON. 181, 184 (1985); Glynn S. Lunney, Jr., Reexamining Copyright’s Incentives-Access Paradigm,
49 VAND. L. REV. 483, 485-86, 492-99 (1996).
2
For representative discussions, see, e.g., Yochai Benkler, An Unhurried View of Private
Ordering in Information Transactions, 53 VAND. L. REV. 2063, 2065-67 (2000); James Boyle, Cruel,
Mean or Lavish? Economic Analysis, Price Discrimination and Digital Intellectual Property, 53 VAND. L.
REV. 2007, 2012-16 (2000); David J. Brennan, Fair Price and Public Goods: A Theory of Value Applied
to Retransmission, 22 INT’L REV. L. & ECON. 347, 349-61 (2002); Fisher, supra note 1, at 1700; Landes
& Posner, supra note 1, at 327, 333; Alfred C. Yen, The Legacy of Feist: Consequences of the Weak
Connection Between Copyright and the Economics of Public Goods, 52 OHIO ST. L.J. 1343, 1364-69
(1991). CHECK L&P. For a useful survey of the literature on public goods, see William H. Oakland,
Theory of Public Goods, in 2 HANDBOOK OF PUBLIC ECONOMICS 485 (Alan J. Auerbach & Martin
Feldstein eds., 1987).
1
that once an author has incurred the fixed costs needed to produce the first
copy of a creative work, it can typically satisfy any number of users at zero
or close to zero marginal cost. Public goods pose a well known economic
conundrum. Basic principles of welfare maximization3 require that all goods
be priced at marginal cost. Under the circumstances described above,
however, pricing a good at marginal cost would prevent the good from being
created, because the product would generate insufficient revenue to cover its
fixed costs. (Indeed, it would be priced at zero and would generate no
revenue at all.) The authors’ need for the means to charge prices that exceed
marginal cost provides a theoretical justification for copyright. On this
account, the reason copyright exists is to provide authors with sufficient
insulation from competition to allow them to recover the fixed costs incurred
to create their works in the first place. The problem is that allowing authors
to engage in supramarginal cost pricing necessarily creates a degree of static
inefficiency by denying excluding some people even though total welfare
would have increased had they been provided with access to the work
The dominant approach, to which I will call the “public goods approach,”
thus casts copyright as an irreconcilable conflict between static and dynamic
efficiency that is not susceptible of a first-best solution. Copyright thus
necessitates resort to a second-best outcome. This basic analytical
framework has come to dominate copyright scholarship. Scholars have
employed it to analyze such varied issues as the fair use doctrine,4
compulsory licenses,5 the breadth of copyright protection,6 and the panoply
of devices that tend to facilitate price discrimination.7
A close examination of the public goods approach reveals that it suffers
from a number of problematic analytical shortcomings. As I will review in
some detail later, its predictions fail to accurately describe a number of
important features of markets for copyrighted works. For example, although
some works clearly generate supracompetitive profits and large market
shares, those features do not arise with nearly the frequency that the
traditional approach to the economics of copyright would suggest.8 Even
more saliently, although public good economics provides a useful way for
3
For purposes of this article, we will follow the standard convention and use total surplus as the
relevant measure of welfare. Although total surplus is not the only available welfare metric, if utility is
linear in the composite commodity, maximization of welfare and maximization of total surplus are
identical. B. Curtis Eaton & Richard G. Lipsey, Product Differentiation, in 1 HANDBOOK OF INDUSTRIAL
ORGANIZATION 723, 729 (Richard Schmalensee & Robert D. Willig eds., 1989).
4
See, e.g., Fisher, supra note 1, at 1705-19; Landes & Posner, supra note 1, at 357-61; Liebowitz,
supra note 1, at 188-92; Lunney, supra note 1, at 546-52.
5
See Brennan, supra note 2, at 349-61.
6
See Landes & Posner, supra note 1, at 347-57; Michael J. Meurer, Copyright Law and Price
Discrimination, 23 CARDOZO L. REV. 55, 109-16 (2001).
7
See Benkler, supra note 2, at 2069-80; Boyle, supra note 2, at 2021-38; Cohen, supra note 1, at
1802-08; William W. Fisher III, Property and Contract on the Internet, 73 CHI.-KENT L. REV. 1203,
1234-40 (1998); Wendy J. Gordon, Intellectual Property as Price Discrimination, 73 CHI.-KENT L. REV.
1367, 1381-86 (1998); Liebowitz, supra note 1, at 192-94; Meurer, supra note 6.
8
See infra Part I.B.2-.3.
2
analyzing static efficiency, it does not quantify dynamic efficiency in a
manner that permits the relevant gains and losses may be traded off against
one another.9
I believe that many of these descriptive and analytical limitations can be
traced to a single assumption implicit in almost all previous analyses, which
is that the relevant products are homogeneous. I would thus like to break
with the established tradition and explore the implications of the somewhat
unorthodox economics of product differentiation for copyright.10 My
analysis reveals that product differentiation theory can provide an alternative
economic justification for why markets for copyrighted works exhibit the
declining-average-cost structure that creates the tension between static and
dynamic efficiency. Indeed, it provides reason to believe that these same
characteristics would arise even with respect to private goods, a conclusion
that raises serious doubts about centrality of the public goods insight. In
addition, product differentiation theory provides an explanation for a wide
variety of market features and analytical problems that the public goods
approach cannot.
Most importantly, product differentiation redresses the central failing of
the public goods approach by providing a method for spanning the gap
between static and dynamic efficiency. It reveals that the irreconcilable
clash between static efficiency and dynamic efficiency that supposedly lies at
the heart of copyright may represent something of a false conflict. What
appears to be static inefficiency from the standpoint of homogeneous
products may in fact represent an optimum when viewed through the lens of
product differentiation. Relaxation of the homogeneity assumption also
reveals the existence of equilibria that maximize dynamic efficiency as well.
Together these insights suggest the somewhat revolutionary possibility that
first-best outcomes are possible under copyright, a conclusion that stands in
stark contrast to the conclusions indicated by public goods theory.
In addition, I would like to offer an extension to current product
differentiation theory that, to my knowledge, are wholly novel. The first is
the manner in which entry by new products tends to reduce deadweight loss
by causing the demand curve to become more elastic.11 This suggests that
steps designed to promote dynamic efficiency by stimulating entry might
also have the indirect effect of promoting static efficiency. Indeed, theory
indicates that such steps are especially appropriate with respect to market
subsegments characterized by high fixed costs and the relative lack of
substitutes. In other words, when confronted with a market subsegment that
is most highly concentrated and susceptible to sustainable profits, the proper
policy response is to increase the amount of surplus captured by firms
9
See infra Part I.B.4.
For surveys of this literature, see J. BEATH & Y. KATSOULACOS, THE ECONOMIC THEORY OF
PRODUCT DIFFERENTIATION (1991); Eaton & Lipsey, supra note 3.
11
See infra Part II.C.2.b.
10
3
operating in that subsegment. This response may at first glance seem
somewhat counterintuitive. It seems less so once one acknowledges the role
of entry in promoting both static and dynamic efficiency when products are
differentiated.
At the same time, the economics of product differentiation raise their own
set of conundrums. In fact, the theory suggests that the strength and breadth
of copyright protection should vary with the availability of substitutes and
the size of the fixed costs.
Although previous scholarship has occasionally contained passing
references to the fact that copyrighted works tend to be differentiated,12 an
extended analysis of the copyright implications of full range of the product
differentiation literature has yet to appear in the literature. My analysis
underscores the regrettable nature of this oversight. It is true that product
differentiation theory yields implications that are extremely contextual and
fact-specific that are not generally susceptible to simple policy inferences.
However, the utility of the insights provided by the differentiated products
analysis more than compensates for its relative lack of tractability.
Part I describes the public goods approach that has become the
established paradigm for the economic analysis of copyright.
The
succeeding parts describe the two major variants of product differentiation
scholarship. Part II examines the branch of the theory that assumes that
consumers have symmetric preferences across all available products,
analyzing both the theory’s superiority to the public goods model in
describing many of the features of markets for copyrighted works and
drawing out the theory’s normative conclusions. Part III considers the
branch of the literature that relaxes the symmetry assumption and allows
consumers’ preferences to vary across brands. Part IV applies the insights
from the foregoing analysis to two of the leading issues in copyright law: the
facilitation of price discrimination and the fair use doctrine.
I. THE TRADITIONAL APPROACH TO THE ECONOMICS OF COPYRIGHT:
PUBLIC GOODS
Section A describes how the basic economics of public goods has shaped
copyright scholarship. Section B reviews the shortcomings and limitations
of the approach. Adherence to the public goods approach gives rise to
12
See Stanley M. Besen & Leo J. Raskind, An Introduction to the Law and Economics of
Intellectual Property, J. ECON. PERSPS., Winter 1991, at 3, 5 n.2; Fisher, supra note 1, at 1702-03; Landes
& Posner, supra note 1, at 327-28 & n.4; Lunney, supra note 1, at 495, 497 n.43; Robert P. Merges,
Intellectual Property Rights and the New Institutional Economics, 53 VAND. L. REV. 1857, 1859 (2000);
Meurer, Copyright Law and Price Discrimination, 23 CARDOZO LAW REVIEW 55, 60-61, 97 (2001). A
recent, unpublished working paper apparently represents the first extended application of product
differentiation theory to copyright. That paper explores only a very limited subsegment of the economic
scholarship on product differentiation, applying a fairly unique model of the asymmetric preferences
branch without any discussion of the insights provided by the symmetric preferences literature or by other
types asymmetric preferences models. See Abramowitz.
4
certain descriptive puzzles that the public goods theory cannot resolve. More
importantly, public good economics fails to provide a welfare metric through
which the static and dynamic efficiency considerations can be balanced.
A. The Second-Best Nature of the Public Goods Equilibrium
The standard approach to analyzing the economics of copyright begins
from the premise that copyrighted works exhibit qualities associated with the
theory of public goods pioneered by Paul Samuelson13 and applied to
intellectual property by Kenneth Arrow.14 Specifically, this approach posits
that consumption of copyrighted works tends to be nonrival, in that
consumption of the work by one person has no effect on the supply of the
work available for consumption by others.15 Thus, unlike rivalrous goods, in
which the demand for scarce factors and other diseconomies of scale in
production eventually cause the marginal cost curve to exhibit the “U” shape
that characterizes most economic analyses, the marginal cost curve for
nonrivalrous goods is more appropriately modeled as flat.
It is the shape of the marginal cost curve that is the source of the unusual
economic problem posed by public goods. When fixed costs are present,
average costs are determined by two components. The first is the
amortization of fixed costs across increasingly large volumes, which tends to
make average cost decrease. The second is marginal cost, which has a direct
impact on unit costs. When the marginal cost curve is “U” shaped, average
cost initially tends to decline, as the decrease in marginal costs and the
amortization of fixed costs reinforce one another. When the average cost
curve is declining it necessarily lies above the marginal cost curve. As
volume increases, however, marginal costs begin to rise and the decline in
average costs associated with the amortization of fixed costs decays
exponentially. At some point, the increase in marginal cost dominates the
13
See Paul A. Samuelson, Aspects of Public Expenditure Theories, 40 REV. ECON. & STATS. 332
(1958); Paul A. Samuelson, Diagrammatic Exposition of a Theory of Public Expenditure, 37 REV. ECON.
& STATS. 350 (1955) [hereinafter Samuelson, Diagrammatic Exposition]; Paul A. Samuelson, The Pure
Theory of Public Expenditure, 36 REV. ECON. & STATS. 387 (1954).
14
Kenneth J. Arrow, Economic Welfare and the Allocation of Resources for Invention, in THE
RATE AND DIRECTION OF INVENTIVE ACTIVITY: ECONOMIC AND SOCIAL FACTORS 609, 616-17 (1962).
15
The definition of a public good often includes the requirement that the good also be
nonexcludable. See, e.g., JOSEPH E. STIGLITZ, ECONOMICS OF THE PUBLIC SECTOR 120 (1988). The
work of Ronald Coase has demonstrated that the lack of excludability does not necessarily lead to market
failure. See R.H. Coase, The Lighthouse in Economics, 17 J.L. & ECON. 357 (1974); R.H. Coase, The
Problem of Social Cost, 3 J.L. & ECON. 1 (1960). As a result, the emerging view is that nonexcludability
should not be regarded as a necessary element of the definition of a public good. See STEPHEN
SHMANSKE, PUBLIC GOODS, MIXED GOODS, AND MONOPOLISTIC COMPETITION 6-7, 16-18 (1991);
Brennan, supra note 2, at 350; J.G. Head, Public Goods and Public Policy, 17 PUB. FINANCE 197, 215
(1962); Oakland, supra note 2, at 486. In addition, technological developments such as digital encryption
are in the process of rendering most copyrighted works fully excludable. See generally Christopher S.
Yoo, A New Economic Model of Television Regulation, 52 EMORY L.J. (forthcoming 2003) (providing a
more extended discussion of this issue).
5
decline in average cost caused by the amortization of fixed costs, and the
marginal cost and average cost curves cross.
So long as the total volume is sufficiently large to support more than one
firm,16 each firm will operate at the point where the average cost curve lies at
or below the marginal cost curve. Eventually the market will reach long-run
equilibrium where the average cost curve is at its minimum, which will also
necessarily be the pint where the average and marginal cost curves intersect.
In such cases, pricing along the marginal cost curve is a viable option,
because the fact that the average cost curve lies at or below the marginal cost
curve guarantees that a firm pricing at marginal cost will generate sufficient
revenue to break even.
A different situation obtains when the marginal cost curve is
nonincreasing. In such cases, the reductions in average cost that arise from
the amortization of fixed costs become inexhaustible, and the marginal cost
never exerts any upward pressure on average costs.17 If that is the case, the
average cost curve will lie above the marginal cost curve over all volumes.
This in turn means that a firm that prices along its marginal cost curve will
generate insufficient revenue to breakeven. If producers of public goods are
expected to generate sufficient revenue to cover the fixed costs needed to
create their works, they must possess sufficient power over price to charge in
excess of marginal cost. This in turn gives rise to what the public goods
approach regards as the economic justification for copyright. Under this
account, copyright exists in order to insulate authors from competition
sufficiently to give them the market power needed to charge the
supramarginal cost prices needed for them to break even.18 If such producers
are assumed to be profit maximizers, they would presumably produce at the
point where the marginal revenue and marginal cost curves intersect.19
16
This proposition does not hold if demand is so low relative to the cost curves that a single firm
could produce the entire industry output more efficiently than could two or more firms. In such cases, the
cost function is said to be subadditive, and the market will tend towards natural monopoly. See WILLIAM
J. BAUMOL ET AL., CONTESTABLE MARKETS AND THE THEORY OF INDUSTRY STRUCTURE 17-21, 169-89
(rev. ed. 1987).
17
Note that this does not require the strong assumption made in the initial discussion of this effect
that marginal cost is zero. See infra text following note 2. This relationship will hold true even with
positive marginal costs so long as the marginal cost is nonincreasing.
18
See, e.g., Boyle, supra note 2, at 2012; Breyer, supra note 1, at 282, 294; Fisher, supra note 1, at
1700; Liebowitz, supra note 1, at 17-18; Lunney, supra note 1, at 494-95, 497.
19
Figure 1 follows what has become the standard approach in copyright scholarship by portraying
the firm producing the public good as confronting a downward-sloping demand curve. See, e.g., Cohen,
supra note 1, at 1802 fig.A, 1804 fig.B; Fisher, supra note 1, at 1701 n.201, 1708 n.232; Fisher, supra
note 7, at 1235 fig.1, 1236 fig.2, 1238 fig.3; Liebowitz, supra note 1, at 18 fig.1. This differs from
Samuelson’s classic formulation, which instead employed marginal rate of substitution curves. See
Samuelson, Diagrammatic Exposition, supra note 13, at 353-54 & chart 5.
Figure 1 also differs from similar diagrams appearing in prior scholarship that portrayed the entire
producer surplus as profit. Such depictions overstate the degree of profit by ignoring the role of fixed
costs. Because of fixed costs, only the portion of the producer surplus lying above the average cost curve
can properly be regarded as profit.
6
Figure 1
Equilibrium Under the Public Goods Approach to Copyright
P ($/unit)
CS
P
Profit
DWL
DWL
Q
MR
AC
MC
D
Q (units)
The problem is that any regime that enables authors to engage in
supramarginal cost pricing necessarily generates static inefficiency by
excluding some would-be customers whose consumption would have caused
total surplus to increase had they been permitted to purchase the work.
Simply put copyright necessarily creates the same type of deadweight loss
associated with monopoly pricing (represented by the light gray triangle
located in the lower right-hand corner under the demand curve). Thus any
solution that promotes dynamic efficiency by providing some ability for
authors to cover their fixed costs inevitably harms static efficiency. Viewed
from this perspective, copyright is inherently a tradeoff between second-best
outcomes, in which no first-best outcome is possible.
A number of scholars have turned to price discrimination as a way to
ameliorate this apparently inherent conflict between static and dynamic
efficiency. Price discrimination has the potential to eliminate the deadweight
losses associated with public goods while simultaneously strengthening the
incentives to create.20 This argument does suffer from one significant
20
Fisher, supra note 1, at 1709-10; Fisher, supra note 7, at 1237-40; Maureen A. O’Rourke,
Copyright Preemption after the ProCD Case: A Market-Based Approach, 12 BERKELEY TECH. L.J. 53,
62, 70-71 (1997).
7
analytic limitation. It has long been recognized that perfect price
discrimination in a market for a public good is unambiguously welfare
enhancing.21 The problem is that perfect price discrimination requires such
an overwhelming amount of information, including every potential
consumers’ reservation price for the good in question along with all of the
substitution effects, that it is in effect impossible.
The best option available to authors as a practical matter must necessarily
be some form of imperfect price discrimination.22
Imperfect price
discrimination falls short of a first-best optimum in two ways. First, the fact
that price discrimination is imperfect necessarily implies that at least some
consumers are likely to remain excluded notwithstanding the fact that their
consumption of the good would cause total surplus to increase. Thus, even
under the best of circumstances, some degree of deadweight loss will likely
remain. Furthermore, although most economists acknowledge that price
discrimination is most likely to cause output to increase and deadweight loss
to fall,23 it is theoretically possible that under certain demand structures
imperfect price discrimination would cause total output to fall, which would
have the effect of increasing the size of deadweight loss. Thus, in contrast to
the situation that obtains with respect to perfect price discrimination, the
welfare implications of imperfect price discrimination are ultimately
ambiguous.
Finally, the market power supposedly created by the copyright and the
ability to engage in price discrimination the inevitable effect of causing the
profits earned by the author to increase by allowing authors to capture a
portion of the consumer surplus (represented by the dotted triangle located in
the upper left-hand corner under the demand curve). According to the
standard account, such profits simply represent a transfer of wealth from
consumers to producers. As such, they have no effect on economic
efficiency,24 although they may have distributional implications.25
21
See James M. Buchanan, Public Goods in Theory and Practice: A Note on the MinasianSamuelson Debate, 10 J.L. & ECON. 193, 195 (1967); Harold Demsetz, The Private Production of Public
Goods, 13 J.L. & ECON. 293, 301-03 (1970); Earl A. Thompson, The Perfectly Competitive Production of
Collective Goods, 50 REV. ECON. STUD. 1, 3-5 (1968).
22
See Meurer, supra note 6, at 69-80.
23
See, e.g., RICHARD G. LIPSEY & PETER O. STEINER, ECONOMICS 252 (6th ed. 1981); F.M.
SCHERER & DAVID ROSS, INDUSTRIAL MARKET STRUCTURE AND ECONOMIC PERFORMANCE 495 (3d ed.
1990).
24
See, e.g., Fisher, supra note 1, at 1702; Lunney, supra note 1, at 497-98. It should be noted that
wealth transfers have no effect on welfare only if one adopts the standard Marshallian device of ignoring
income effects. Even when income effects are present, the percentage errors involved in taking areas
under Marshallian demand curves are likely to be relatively small. See Michael Spence & Bruce Owen,
Television Programming, Monopolistic Competition, and Welfare, 91 Q.J. ECON. 103, 104 n.4 (1977).
For an interesting discussion of how income effects can give wealth transfers greater economic
significance, see HAL VARIAN, MICROECONOMIC ANALYSIS 160-68 (3d ed. 1984).
25
See Boyle, supra note 2, at 2025-26, 2029; Cohen, supra note 1, at 1806. Conversely, William
Fisher argues that such wealth transfers can actually promote distributional equity by providing the
justification for authors to accept adjustments in other areas that reduce the scope of copyright protection.
Fisher, supra note 7, at 1239-40.
8
B. Puzzles that the Public Goods Approach Cannot Resolve
The central economic feature driving the public goods analysis is the
presence of constantly declining average costs. It is this feature that thus
transforms copyright into an intractable pricing problem that is not
susceptible of a first-best solution.
Before accepting the portrayal of copyright as an irreconcilable conflict
between static and dynamic efficiency, it is worth evaluating the public
goods approach against the generally accepted standard for judging an
economic theory’s validity, i.e., the ability to make predictions that conform
with empirical reality.26 A critical examination of the public goods approach
yields a number of implications that do not appear to correspond with
observed markets for copyrighted works. This lack of fit with real-world
facts suggests that the time may well be ripe for a new theory.
1. The Questionable Assumption of Constant Marginal Costs
One criticism of the public goods approach focuses on its assumption that
marginal costs are constant. Indeed, it is this premise that causes average
cost to decline across all volumes, which in turn drives the key aspects of the
public goods equilibrium. The problem with this premise is that copyrights
are typically combined with other inputs before they are sold to end users.
Because those other inputs will typically exhibit rivalry in consumption, it is
far from clear that the final good actually marketed to consumers will exhibit
the nonincreasing marginal costs needed to give rise to the supposedly
intractable pricing problem identified by public good economics.27
2. The Absence of Natural Monopoly Despite Constantly Declining
Average Costs
Given the assumption that average costs will decline across all volumes,
one would expect that markets for public goods exhibit inexhaustible
economies of scale. As a result, theory suggests that markets for public
goods should display a strong tendency towards natural monopoly, since the
producer with the largest volume would enjoy cost advantages that should
allow it to drive all of its competitors out of the market.28 Because the public
goods approach posits that every copyrighted work manifests declining
26
See, e.g., MILTON FRIEDMAN, The Methodology of Positive Economics, in ESSAYS IN POSITIVE
ECONOMICS 3, 8-9 (1953); GEORGE J. STIGLER, Monopolistic Competition in Retrospect, in FIVE
LECTURES ON ECONOMIC PROBLEMS 12, 23 (1949), reprinted in GEORGE J. STIGLER, THE
ORGANIZATION OF INDUSTRY 309 (1968); G.C. Archibald, Chamberlin versus Chicago, 29 REV. ECON.
STUDS. 2, 2-5 (1961).
27
Edmund W. Kitch, Elementary and Persistent Errors in the Economic Analysis of Intellectual
Property, 53 VAND. L. REV. 1727, 1733-34 (2000). For examples of this effect in the context of
television programming, see Demsetz, supra note 21, at 296-303; Yoo, supra note 15, at _.
28
See WILLIAM W. SHARKEY, THE THEORY OF NATURAL MONOPOLY 47 (1982).
9
average costs, one would expect natural monopoly to be endemic. The
tendency towards natural monopoly also threatens to give the market a
winner-take-all quality that promotes the type of destructive races to get to
market has commanded so much attention in patent law.29 It also dictates
that any attempt by more than producer to create the work will represent a
waste of resources.30
Unfortunately, these conclusions does not appear to jibe with empirical
reality. The most casual perusal of the market for copyrighted material
reveals that most, if not all, of them appear to face sustainable competition
from a multitude of other producers. To be sure, many copyrighted works
undoubtedly appear to enjoy dominant market shares. The problem is that
public goods theory would suggest that all, and not just some, copyrighted
works would exhibit this quality. The puzzle from the standpoint of public
good economics, then, is not why some works have dominant market shares,
but rather why so many do not.
3. The Persistence of Market Power Notwithstanding the Prevalence
of Substitutes and the Absence of Supracompetitive Profits
As noted earlier,31 the public goods approach posits that copyright exists
in order to provide authors with power over price. Without such market
power, they would be unable to impose the supramarginal cost pricing that is
required if the works in question are to be created. Although it was once
commonly assumed that copyrights automatically conveyed monopoly
power,32 it is now widely acknowledged that that is only rarely the case.
Instead, copyright is more properly regarded simply as a grant of an
exclusive right. Granting one person the exclusive right to control any piece
of property does not necessarily give that person any market power.
Whether or not it conveys market power depends largely on the availability
of substitutes. For example, the fact that homeowners have the exclusive
right to use their houses does not give them the power to charge
supracompetitive prices. Such power over price only arises if there are no
other houses available for purchase.
Viewed from this perspective, one would expect that the ready
availability of substitutes would cause any extant market power to dissipate.
29
See Meurer, supra note 6, at 97. For an overview of the literature on “patent races,” see Jennifer
F. Reinganum, The Timing of Innovation: Research, Development, and Diffusion, in 1 HANDBOOK OF
INDUSTRIAL ORGANIZATION, supra note 3, at 753.
30
See C. Edwin Baker, Giving the Audience What It Wants, 58 OHIO ST. L.J. 311, 339-40 (1997).
This is analogous to concerns about “overbuilding” in natural monopoly industries, such as cable
television. The argument is that if a market is destined to collapse into natural monopoly, allowing more
than one firm to compete for the market would only waste resources. For a critical overview, see Hazlett,
supra note 48, at _.
31
See supra note 18 and accompanying text.
32
See, e.g., United States v. Loew’s, Inc., 371 U.S. 38, 45 (1962); United States v.
Paramount.Pictures, Inc., 334 U.S. 131 (1948).
10
The so-called “idea-expression dichotomy” provides that copyright does not
protect “any idea, procedure, process, system, method of operation, concept,
principle or discovery” contained in a work.33 As a result, nothing in
copyright law forecloses competitors from creating works with the same
functional characteristics as any copyrighted work.34
These facts would lead one to expect that the widespread availability of
substitutes would vitiate copyright holders’ ability to charge prices that
exceed marginal cost. If that were the case, products would not be able to
generate sufficient revenue to cover their fixed costs and should not be able
to exist. Yet that clearly the presence of substitutes has not prevented
authors from charging prices sufficient to cover their fixed costs.35 The
coexistence of market power and substitutes thus represents something of a
paradox under the public goods approach to copyright. In fact, both
proponents and critics of strengthening copyright protection have recognized
as much without offering any way to resolve the quandary.36
A similar problem is posed by the simultaneous coexistence of market
power and a wide array of copyrighted works that generate no
supracompetitive profits. Under the public goods theory, copyright exists to
give authors sufficient power over price to recover the fixed costs of
producing the first copy of their work. If successful, public goods theory
predicts that that same market power ought to enable those authors to earn
supracompetitive profits. Again, there is little doubt that some copyrighted
works have been able to command what are almost unquestionably
supracompetitive returns. Public goods theory, however, predicts that every
product that is produced ought to generate such profits. Yet public goods
theory fails to provide an adequate explanation of why such profits would
not be universal.37
33
17 U.S.C. § 102(b); accord Eldred v. Ashcroft, 123 S. Ct. 769, 788-89 (2003); Feist Publ’ns, Inc.
v. Rural Tel. Serv. Co., 499 U.S. 340, 349-50 (1991); Harper & Row Publishers, Inc. v. Nation Enters.,
471 U.S. 539, 547-48, 556 (1985); N.Y. Times Co. v. United States, 403 U.S. 713, 726 n.* (1971)
(Brennan, J., concurring); Int’l News Service v. Associated Press, 248 U.S. 215, 234 (1918); Baker v.
Selden, 101 U.S. 99, 104 (1879).
34
See Kitch, supra note 27, at 1730, 1734..
35
Several early copyright theorists proposed that delays in copying and other sources of market
friction would provide authors with a sufficient window to recoup their fixed costs. See Breyer, supra
note 1, at 299-301; Robert M. Hurt & Robert M. Schuchman, The Economic Rationale of Copyright, AM.
ECON. REV., May 1966, at 421, 427-28; A. Plant, the Economic Aspects of Copyright in Books, 1
ECONOMICA 167 (1934). Those arguments were not even convincing in their day. See Barry W.
Tyerman, The Economic Rationale for Copyright Protection for Published Books: A Reply to Professor
Breyer, 18 UCLA L. REV. 1100 (1971); see also Fisher, supra note 1, at 1708 n.231 (calling Tyerman’s
refutation of Breyer “convincing”). They are even less so today, where the costs of copying and
distribution have plummeted.
36
See Boyle, supra note 2, at 2028-29, 2032, 2037-38; Kitch, supra note 27, at 1734-35.
37
It is conceivable that risk might explain the prevalence of copyrighted works that do not earn
profits. Specifically, authors who face uncertainty ex ante about whether their work will generate
supracompetitive returns may undertake the investments needed to create the work only to find out ex post
that the particular work in question will not generate any such benefits. After the work has been created,
however, the author may find it most beneficial to sell the work at competitive prices. Authors who
confront such uncertainty should be expected to base their decisions based on the expected value of the
11
4. The Inability to Resolve the Tradeoff Between Static and Dynamic
Efficiency
Perhaps the most significant analytical shortcoming of the public goods
approach is its inability to provide a common welfare metric for balancing
the countervailing static and dynamic efficiency considerations. The public
goods approach focuses almost exclusively on the most efficient allocation
of goods that have already been produced, adopting the conventional
economic approach endorsing of allocating the good by pricing it at marginal
cost. As a result, it does not provide a basis for determining how many
resources should be devoted to producing the public goods in the first
place.38
Some scholars have flatly asserted a preference for dynamic efficiency
considerations over static efficiency considerations on the grounds that when
compounded over time, the long-run benefits in dynamic efficiency will
dominate whatever short-run static inefficiency losses that may exist.39 Even
this assertion, however, ultimately begs the question. While clearly favoring
dynamic efficiency over static efficiency, it fails to provide any way to
determine what set of policies will in fact optimize dynamic efficiency.
Although some scholars have turned to such rough metrics as an
“incentive/loss ratio” to provide some indication of whether the net
economic benefits.40 Even the proponents of this approach do not pretend
that it provides anything more than a rough metric that did little formalize
questions of how much creative activity was optimal.
* * *
The public goods approach to copyright would thus appear to suffer from
several critical analytical deficiencies. The theory leads to predictions that
do not conform well to empirical reality. Even more glaringly, although it
provides a way to formalize the static efficiency implications of copyright, it
fails to provide a basis for evaluating dynamic efficiency, let alone provide a
common metric through which both considerations can be balanced against
one another. I now turn to the insights generated by product differentiation
theory as a means for resolving each of these analytical conundrums.
work discounted by the extent to which they are risk averse. Unless the bulk of the expected returns are
comprised of a very small number of extremely high-value works, this effect would not ordinarily lead
one to expect the market to produce the number of unprofitable works observed.
38
Jora R. Minasian, Television Pricing and the Theory of Public Goods, 7 J.L. & ECON. 71, 73, 79
(1964).
39
Brennan, supra note 2, at 355 (citing Janusz Ordover & William Baumol, Antitrust Policy and
High-Technology Industries, 4 OXFORD REV ECON. POL’Y 13, 32 (1988)).
40
Fisher, supra note 1, at 1707-19 (using the ratio of producer surplus to deadweight loss as a
rough measure the cost-effectiveness of social losses). Fisher adapted the basic approach from Louis
Kaplow, The Patent-Antitrust Intersection: A Reappraisal, 97 HARV. L. REV. 1813, 1829-34 & n.54
(1984).
12
II. THE DIFFERENTIATED PRODUCTS APPROACH TO COPYRIGHT:
SYMMETRIC PREFERENCE MODELS
The literature on the economics of product differentiation can be divided
into two distinct approaches. The first is the symmetric preferences branch,
which centers around the theory of “monopolistic competition” first
elucidated by Edward Chamberlin.41
The other is the asymmetric
preferences branch, which draws inspiration largely from the critique of
monopolistic competition offered by Nicholas Kaldor42 and the spatial
competition models pioneered by Harold Hotelling.43
My analysis begins by drawing primarily from the symmetric preferences
literature. Section A describes the nature of the equilibrium under
monopolistic competition. Section B evaluates the theory’s ability to explain
the descriptive and analytical puzzles identified above that the public goods
approach is unable to resolve. Section C analyzes the welfare implications of
monopolistically competitive equilibrium. Section D responds to some of
the criticisms that have been leveled at monopolistic competition theory in
the past.
I begin with monopolistically competition theory because it offers certain
analytical advantages. Specifically, because monopolistic competition
portrays market interactions in the classic price-quantity space, it is quite
easily integrated into a conventional welfare analysis. The primary downside
of monopolistic competition is that it does not model product diversity
directly. As a result, I will turn in the next Part to asymmetric preferences
models, which have the advantage of representing product differentiation
directly by allowing firms to locate themselves along a spectrum of product
attributes. While modeling product differentiation in an attribute space
makes many of the aspects of product differentiation more intuitive, it raises
the converse problem of not being easily adapted into a direct analysis of
economic surplus.
A. The Theory of Monopolistic Competition
Monopolistic competition retains the key assumptions of perfect
competition while relaxing one key assumption.44
Under perfect
competition, the goods produced by firms competing within an industry act
as perfect substitutes for one another. As a result, no firm can raise its prices
41
See EDWARD HASTINGS CHAMBERLIN, THE THEORY OF MONOPOLISTIC COMPETITION (7th ed.
1956) (1933).
42
Nicholas Kaldor, Market Imperfection and Excess Capacity, 2 ECONOMICA 33 (1935).
43
Harold Hotelling, Stability in Competition, 34 ECON. J. 41 (1929).
44
For formal statements of the assumptions of monopolistic competition, see BEATH &
KATSOULACOS, supra note 10, at _; Oliver D. Hart, Monopolistic Competition in the Spirit of
Chamberlin: A General Model, 52 REV. ECON. STUDS. 529, 529 (1985); CHAMBERLIN, supra note 41, at
_; Kaldor, supra note 42, at 34-37.
13
above the equilibrium price without having all of its customers transfer their
purchases to its competitors. This makes each firm a price taker with no
power over price.
Chamberlin relaxed the assumption that competing products were
homogeneous. Instead he assumed that each firm in the industry sold a
different product that served as imperfect substitutes for every other product
manufactured by their competitors. Allowing products to serve as imperfect
substitutes for one another made it theoretically possible for a firm to raise
price without losing all of its customers, since the firm would be able to
retain those customers who valued their particular version of the good
particularly highly. Product differentiation thus provides each firm sufficient
power over price to justify modeling the economics as if each firm faced a
downward-sloping demand curve.
To simplify his analysis, Chamberlin posited that each participating firm
faced identical cost and demand curves. This allowed him to employ a
single graph portraying the price-quantity response of a representative firm
to model the entire market. The fact that monopolistic competition creates
its model in terms of a representative firm rather than at the industry level
has some important implications that bear emphasizing. When a similar
price-quantity analysis is applied to a monopoly case, the only relevant
criterion is the maximization of surplus depicted on that particular graph.
Because the market is a monopoly, there is no other potential source of
welfare in that industry. The welfare calculus becomes somewhat more
complex when the market is analyzed in terms of a representative firm. In
that case, total welfare depends on the number of works created as well as on
the consumer and producer surplus generated by any particular work. In
other words, any reduction in the surplus generated on any particular graph
has the potential of being offset by an increase in the total number of graphs.
In this way, monopolistic competition is able to capture an element of
dynamic efficiency that is unavailable under the public goods approach to
copyright.
Chamberlin’s most important assumption was that consumers’
preferences were symmetric with respect to each product in the industry
group. Stated more formally, this assumption posits that the cross-price
elasticities of demand are equal with respect to all other works.45 The
primary effect of this assumption is to place each product in equal
competition with all other products rather than in localized competition with
a smaller set of near neighbors. Because a new product will steal business
equally from all of incumbent producers, the effects of new entry to spread
uniformly across the entire industry.
A profit-maximizing firm will produce where the marginal revenue and
marginal cost curves intersect. The fact that the demand curve is downward
45
Kaldor, supra note 42, at 35 n.2.
14
sloping causes the marginal revenue curve to lie below the demand curve.
Having identified the profit maximizing level of production, the firm will
charge the maximum that it can receive for that quantity, which is
represented by where that quantity falls along the demand curve faced by
that firm. Because this necessarily means that monopolistically competitive
firms will price above marginal cost, some degree of deadweight loss is
inevitable.
Figure 2
Short-Run and Long-Run Equilibrium Under Monopolistic Competition
P ($/unit)
t
Profi
MC
PSR
AC
ProfitSR
PLR
DWLLR
DLR
QLR
QSR
MRSR
DSR
Q (units)
Should the equilibrium price exceed average cost, it is possible that a firm
operating in a monopolistically competitive market might earn short-run
profits. Any such profits should not prove enduring over the long run. This
is because monopolistic competition retains the assumption from perfect
competition that entry is free. The existence of supracompetitive profits will
attract entry by firms selling similar products. Because all of the competing
products are in equal competition with one another, entry by a new product
will steal business equally from each of the incumbent firms. Entry by these
new products will cause the demand curve confronting each of the
incumbents to shift backwards, as customers substitute purchases of the new
15
product for the incumbents’. The increased availability of substitutes should
also cause the demand curve to flatten. Such entry will continue until no
supracompetitive profits remain, at which point the firm’s demand curve will
be tangent to its average cost curve.
B. The Descriptive Advantages of Applying Monopolistic Competition
to Copyright
The monopolistically competitive equilibrium has a number of interesting
features. Most importantly for our purposes, it provides an explanation for
each of the analytical anomalies and deficiencies identified in Part I.B.
1. The Irrelevance of Nonrivalry
One of the most interesting aspects of the monopolistically competitive
equilibrium is the shape and relative position of the average and marginal
cost curves. Because the principle of diminishing marginal returns dictates
that demand curves be downward sloping, equilibrium necessarily occurs
where the average cost curve is downward sloping and above marginal cost.
The shape and relative position of the average and marginal cost curves over
the relevant volumes are thus remarkably similar to those customarily used to
model public goods. In fact, many of the leading monopolistic competition
analyses model average and marginal costs in precisely this manner.46
This analysis suggests that the relative position of cost curves that gave
rise to fundamental pricing problem that represents the central focus of
copyright under the public goods approach may not be the result of the lack
of rivalry in consumption. On the contrary, monopolistic competition theory
suggests that the cost curves will partake of the same essential features even
if the goods in question were completely rivalrous and both the average and
marginal cost curves take on the customary “U” shape.
This equilibrium characteristic allows monopolistic competition to
explain why works continue to exhibit qualities associated with public goods
even though the copyrighted material must typically be combined with other,
rivalrous inputs before it can be sold to consumers.47 Indeed, it suggests that
the tension between static and dynamic efficiency that lies at the heart of
copyright would exist even if the goods involved were completely private
goods. Although the lack of rivalry would clearly reinforce the same effect,
46
See, e.g., BEATH & KATSOULACOS, supra note 10, at 50; Avinash K. Dixit & Joseph E. Stiglitz,
Monopolistic Competition and Optimum Product Diversity, 67 AM. ECON. REV. 297, 299 (1977); Eaton &
Lipsey, supra note 3, at 729; Michael Spence, Product Differentiation and Welfare, 66 AM. ECON. REV.
407, 409 fig.1, 411 (1976). The discussions of monopolistic competition in many leading industrial
organization textbooks also make the same assumption. See, e.g., DENNIS W. CARLTON & JEFFREY M.
PERLOFF, MODERN INDUSTRIAL ORGANIZATION 202-07 & figs.7.1, 7.2 (3d ed. 2000); JEFFREY CHURCH
& ROGER WARE, INDUSTRIAL ORGANIZATION 377-78 & figs.11.3-11.4 (2000).
47
See supra Part I.B.1.
16
it is not a necessary condition for the pricing problem underlying copyright
to arise.
2. Product Differentiation as a Means for Averting Natural
Monopoly
Monopolistic competition theory also provides an explanation for why
markets for copyrighted works tend not to devolve into natural monopolies
as would be predicted under the public goods approach. Although declining
costs can give rise to a winner-take-all type of competition when products
are homogeneous, a different situation obtains when products are
differentiated. The fact that products serve as imperfect substitutes for one
another allows multiple producers to coexist notwithstanding the presence of
unexhausted economies of scale, as evidenced by the fact that the market
reaches a multiple-producer equilibrium even though each firm’s average
cost curve is declining. Variations in customer preferences thus make it
possible for multiple declining-cost firms to coexist by targeting different
segments of the overall customer base.48
Product differentiation also allows multiple providers to incur the fixed
costs needed to produce similar works without wasting resources. Indeed,
allowing for the possibility of differentiated products makes investments that
look redundant from the standpoint of homogeneous products a necessary
prerequisite for the maximization of welfare. Product differentiation also
helps defray worries that such investments will lead to destructive races to
get to market.49 The possibility that a firm may compete on some other basis
than cost opens up the possibility of subsequent entry even though the new
entrant would suffer from cost and volume disadvantages that would be
insuperable if products were homogeneous.
48
See BRUCE M. OWEN, THE INTERNET CHALLENGE TO TELEVISION 32 (1999); Thomas W.
Hazlett, Private Monopoly and the Public Interest: An Economic Analysis of the Cable Television
Franchise, 134 U. PA. L. REV. 1335, 1355-56, 1368-70 (1986). This is the supply-side analog to the
manner in which customer heterogeneity can mitigate the demand-side economies of scale resulting from
network economic effects. See WILLIAM F. BAXTER ET AL., RETAIL BANKING IN THE ELECTRONIC AGE
101-20 (1977); Joseph Farrell & Garth Saloner, Standardization and Variety, 20 ECON. LETTERS 71
(1986); Michael L. Katz & Carl Shapiro, Systems Competition and Network Effects, 8 J. ECON. PERSP. 93,
106 (1994); S.J. Liebowitz & Stephen E. Margolis, Should Technology Choice Be a Concern of Antitrust
Policy?, 9 HARV. J.L. & TECH. 283, 292 (1996). See generally Christopher S. Yoo, Vertical Integration
and Media Regulation in the New Economy 19 YALE J. ON REG. 171, 272, 280-81 (2002) (reviewing this
literature); Daniel F. Spulber & Christopher S. Yoo, Access to Networks: Economic and Constitutional
Connections, CORNELL L. REV. (forthcoming May 2003) (same).
49
See Meurer, supra note 6, at 97. For an overview of the literature on “patent races,” see Jennifer
F. Reinganum, The Timing of Innovation: Research, Development, and Diffusion, in 1 HANDBOOK OF
INDUSTRIAL ORGANIZATION, supra note 3, at 753.
17
3. The Coexistence of Market Power, Free Entry, and the Absence of
Supracompetitive Profits
Product differentiation theory also provides an explanation of how firms
can retain sufficient market power to charge above marginal cost and engage
in price discrimination despite the ready availability of substitutes.50 In the
case of monopolistic competition, the requisite power over price comes from
imperfections in substitutability that by definition cannot be completely
dissipated by the availability of substitutes. As a result, every author will
retain a degree market power notwithstanding the fact that other authors
remain free to create works with the same functional characteristics. Were it
otherwise, competition would eliminate authors’ ability to engage in price
discrimination and to generate sufficient revenues to cover the fixed costs to
produce the work in the first place.
This analysis calls into question the conclusion advanced under the public
goods theory that identifies copyright as the source of the market power
authors need to finance production of their works. Monopolistic competition
theory indicates that such market power comes more from the product
differentiation inherent in the underlying market than from any supposed
monopoly power granted by copyright.
This reality is eloquently
demonstrated by the fact that some creative works that lacked any
intellectual property protection whatsoever have nonetheless been able to
generate sufficient revenue to cover fixed costs.
Although the possibility of free entry may be insufficient to dissipate
market power, it may serve to dissipate any short-run profits that may exist.
Put another way, entry causes any profits that arise by virtue of the power
over price to be rebated back to consumers in the form of increased product
variety. Monopolistic competition theory thus serves to resolve another
paradox under the public goods approach, which is how market power can
exist without allowing most firms to earn supracompetitive returns.
This revelation highlights the importance of analyzing markets in terms
other than just the two poles of perfect competition on the one hand and pure
monopoly on the other hand. It is the resort to an intermediate model of
imperfect competition such as monopolistic competition that severs the
apparent tie between market power and supracompetitive profits that seems
inevitable if the only options entertained are perfect competition and
monopoly.51
50
See Severin Borenstein, Price Discrimination in Free-Entry Markets, 16 RAND J. ECON. 380,
381, 394 (1985); Benjamin Klein, Market Power in Antitrust: Economic Analysis after Kodak, 3 SUP. CT.
ECON. REV. 43, 74-78 & n.60 (1993). Note that the simultaneous coexistence of these factors ultimately
depends on some degree of preference asymmetry of the type discussed in Part III. See Borenstein, supra,
at 387.
51
See, e.g., Boyle, supra note 2, at 2025-27.
18
Finally, product differentiation theory also largely renders moot the
objection that strengthening copyright protection and facilitating price
discrimination raises distributional concerns,52 as the economics of product
differentiation suggest that in most cases no such profits will exist. This
conclusion is, however, subject to a caveat. The fact that fixed costs are
often indivisible gives rise to the well-known “integer problem” in which n
firms might earn small profits while n +1 firms would run losses. In such
cases, the equilibrium would consist of n firms each earning some degree of
sustainable profits.53 In “large economies” (i.e., if n is relatively large), such
profits will be relatively small. Whether a economy is “large” in this manner
does not depend upon the absolute size of the fixed costs involved. Instead,
it depends on the magnitude of the fixed costs relative to the overall
market.54
It should be noted, however, that this (near) zero-profit result relies
heavily on the assumption that consumer preferences are symmetric and that
all products are in equal competition with one another. We will eventually
see that relaxation of this assumption makes the danger of sustainable profits
far more acute.55
C. The Welfare Implications of Monopolistic Competition
Product differentiation theory thus provides a way to analyze the
economics of copyright that is better able to explain many of the unique
features of markets for creative works.
In addition, monopolistic
competition theory provides a whole new perspective on the welfare analysis
of copyright. Monopolistic competition transforms the welfare analysis in
three ways. First, it provides a new way to interpret the static efficiency that
suggests that what appears to be welfare losses from the standpoint of
homogenous products may not be welfare losses at all. Second, it surpasses
public goods theory by providing a way to formalize the dynamic efficiency
side of the welfare calculus while providing a common metric through which
static and dynamic efficiency considerations can be traded off. Third, it
sheds new insights into the role that fixed costs play in a more dynamic
economic environment
52
See supra note 25 and accompanying text.
The seminal statement was offered by Nicholas Kaldor. See Kaldor, supra note 42, at 42-43.
54
BEATH & KATSOULACOS, supra note 10, at _; Oliver D. Hart, Monopolistic Competition in a
Large Economy with Differentiated Commodities, 46 REV. ECON. STUD. 1, 1-2 (1979); Larry E. Jones,
The Efficiency of Monopolistically Competition Equilibria in Large Economies:
Commodity
Differentiation with Gross Substitutes, 41 J. ECON. THEORY 356 (1987); see also B. Curtis Eaton &
Myrna Holtz Wooders, Sophisticated Entry in a Model of Spatial Competition, 16 RAND J. ECON. 282
(1985) (deriving similar results in an asymmetric preferences model).
55
See infra Part III.A.
53
19
1. A New Perspective on Static Efficiency
Under the conventional approach, the fact that monopolistically
competitive markets reach equilibrium on the downward-sloping portion of
the average cost curve would appear to lead unambiguously to some degree
of static inefficiency. Because in equilibrium average cost exceeds marginal
cost, any firm that is charging enough to break even must necessarily be
setting prices above marginal cost, which inevitably leads to deadweight
loss. In addition, the fact that monopolistically competitive markets do not
reach long-run equilibrium at the point that minimizes average cost led
Chamberlin to conclude that these markets necessarily operate with “excess
capacity.”56 This suggests that unrealized gains from trade existed if firms
could only find a way to agree. Price discrimination would likely help
alleviate these problems, but given that any such regime would necessarily
be imperfect, not completely so.
Later theorists pointed out that, given the ready availability of substitutes,
any such deadweight losses are likely to be small.57 Even more importantly,
they realized that these initial conclusions were too simplistic and failed to
take into account the full implications of product differentiation. The more
refined approaches to monopolistic competition recognize that the supposed
shortfall in static efficiency may in fact be illusory.
This is because when products are undifferentiated, firms can compete
only on a single dimension: price. Limiting competition to a single
dimension also greatly simplifies the relevant welfare analysis by limiting it
to the aggregate spread between reservation prices and actual prices (i.e.,
total surplus). The welfare calculus changes dramatically when the
competing firms sell differentiated products. In such markets, firms compete
not only by offering cheaper prices, but also by offering products with
attributes that come closer to particular customers’ ideal combination of
product attributes. The multidimensionality of this competition makes
simple price-cost comparisons an inapt way to determine social welfare,
since such an approach would fail to take into account the welfare that is
created when buyers are able to consume products that lie closer to their
ideal preferences. This is an aspect of economic welfare that is not captured
by deadweight loss, excess capacity, or any other equilibrium feature that can
56
See, e.g., CHAMBERLIN, supra note 41, at 104-10. The “excess capacity” result provides still
another way to distinguish the differentiated products approach to copyright from the public goods
approach. Because the cost curves for monopolistically competitive firms are “U” shaped, producing at
minimum average cost is feasible. In contrast, the average cost curves under the public goods approach
are constantly decreasing, thereby making it impossible for the firm to operate at the level that minimizes
average cost. Furthermore, public goods’ tendency towards natural monopoly eliminates any possibility
of gains from trade resulting from having other firms reduce their production.
57
See BEATH & KATSOULACOS, supra note 10, at _; ROBERT S. PINDYCK & DANIEL L.
RUBINFELD, MICROECONOMICS 438 (4th ed. 1998).
20
be represented in the price-quantity space employed under traditional
economics.
As a result, product differentiation raises the possibility that any
deadweight losses might be offset by welfare gains resulting from product
variety.58 The same can be said for the welfare losses associated with the
fact that equilibrium is reached on the downward-sloping portion of the
average cost curve.59 In fact, if the dimension upon which products are
differentiated is added and the market represented in three dimensions, it is
quite possible that the resulting equilibrium is a true three-dimensional
optimum.60
2. The Formalization of Dynamic Efficiency
a. Appropriability as a Determinant of Dynamic Efficiency
In addition to suggesting that the deadweight losses identified by public
goods theory as problematic may not actually represent static inefficiency,
product differentiation theory also surpasses public goods theory by offering
a simple welfare-maximization principle to evaluate dynamic efficiency.
This principle simply states that a work should be produced whenever the
surplus it would create exceeds the costs needed to produce it.61 Stated in
terms of the graphical representation, this condition is met whenever the area
under the demand curve and above the marginal cost curve is larger than the
fixed costs.
The problem is that any single-part pricing regime will necessarily allow
the producer to appropriate only a fraction of the available surplus. Indeed,
if the good is priced at marginal cost, the producer will not be able to capture
any surplus whatsoever, and no goods will be produced. Even if the firm is
allowed to price so as to maximize profits, it will only be able to capture a
fraction of the available surplus. For example, in the classic case of a firm
facing a linear demand firm will only be able to capture fifty percent of the
available surplus. Although that would permit the creation of works whose
fixed costs comprise less than fifty percent of the available surplus, works
requiring fixed costs that exceed fifty percent of the total surplus would still
not be created notwithstanding the fact that their creation would increase
total surplus.
58
See Eaton & Lipsey, supra note 3, at 742; Spence, supra note 46, at 411.
See BEATH & KATSOULACOS, supra note 10, at _; Harold Demsetz, The Nature of Equilibrium in
Monopolistic Competition, 67 J. POL. ECON. 21, 22 (1959); Dixit & Stiglitz, supra note 46, at 301-02.
60
BEATH & KATSOULACOS, supra note 10, at _.
61
Spence, supra note 46, at 407-08; Michael Spence, Product Selection, Fixed Costs, and
Monopolistic Competition, 43 REV. ECON. STUD. 217, 218-20, 224, 230; (1976); Spence & Owen, supra
note 24, at 110-11; see also Dixit & Stiglitz, supra note 46, at 297; Roger W. Koenker & Martin K. Perry,
Product Differentiation, Monopolistic Competition, and Public Policy, 12 BELL J. ECON. 217, 226 (1981).
59
21
Indeed, one can imagine a range of products requiring fixed costs that
range anywhere from fifty percent up to one hundred percent of the total
surplus, each of which would cause total surplus to increase. If producing
firms were able to appropriate all of the surplus created by their products,
they would be able to use the entirety of that surplus to fund the fixed costs,
and all welfare-enhancing works would be made. If authors are able to
capture only a fraction of the benefits created by their works, some welfareenhancing programs will not be made. The greater the slippage in the firm’s
ability to capture all of the available surplus, the more welfare-enhancing
works will be lost.62
Figure 3
Appropriability as a Function of Inverse Demand
P ($/unit)
P ($/unit)
PPC
PPC
PS
PS
D
D
Q (units)
Q (units)
62
Some scholars have argued that the degree of appropriability should be calibrated on a case-bycase basis so that it provides only the minimum return necessary to fund creation of the work. See, e.g.,
Dan L. Burk, Muddy Rules for Cyberspace, 21 CARDOZO L. REV. 121, 133-34 (1999); Lawrence Lessig,
The Law of the Horse: What Cyberlaw Might Teach, 113 HARV. L. REV. 501, 527 (1999). Monopolistic
competition theorists find the informational requirements to implement any such system unmanageable,
since it would in effect require the government to know the reservation price of each individual for each
work and all of the substitution effects. Spence & Owen, supra note 24, at 122.
22
Monopolistic competition theorists were also able to use this principle to
identify those products against which the market will be biased.63 For
example, the market will be biased against products with high fixed costs,
since those works will require a higher degree of appropriability if they are to
be created. In addition, the market will be biased against products whose
demands are structured such that permits appropriation of a relatively low
proportion of the total surplus. These are products with steep inverse
demand functions (such as the one represented in the left-hand graph in
Figure 364), which is a concept related to, but not identical with, own-price
elasticity. These products tend to be those whose benefits are concentrated
in a relatively small group of consumers who value the product particularly
highly.
As noted earlier,65 appropriability of surplus is not a concern in
conventional markets that operate on the non-decreasing portion of the
average cost curve. In such markets, average cost lies at or below the
marginal cost curve. Fixed costs in such markets are self financing in
equilibrium and need not be funded out of a transfer of surplus from
consumers to producers. A different situation obtains under monopolistic
competition. The fact that the average cost curve is decreasing throughout
the relevant range of output all but dictates that the marginal cost curve is
either flat or decreasing.66 As a result, if the good were priced at marginal
cost, there would be no producer surplus whatsoever from which to finance
fixed costs. The revenue to cover fixed costs thus must necessarily come
from a transfer of surplus from consumers to producers.
This also represents a fairly sharp departure from the approach to
appropriability taken by public good economics, as exemplified by the
position taken with respect to price discrimination. As noted earlier,67 the
public goods approach found the welfare implications of price discrimination
to be ambiguous. Imperfect price discrimination would unambiguously
cause profits to increase. Because this simply effected a transfer of wealth,
public goods theory regarded it as irrelevant from the standpoint of welfare
maximization. More important from the standpoint of the public goods
63
See BEATH & KATSOULACOS, supra note 10, at 59; Dixit & Stiglitz, supra note 46, at 307-08;
Koenker & Perry, supra note 61, at 226; Spence, supra note 46, at 408, 413; Spence, supra note 61, at
224, 234; Spence & Owen, supra note 24, at 111-12.
64
Figure 3 is adapted from Spence, supra note 46, at 409 fig.1; and BRUCE M. OWEN & STEVEN S.
WILDMAN, VIDEO ECONOMICS 112 figs.4.5 & 4.6 (1992).
65
See supra note 16 and accompanying text.
66
It is theoretically possible that a monopolistically competitive firm could reach equilibrium over
the interval just before the average and marginal cost curves cross where the marginal cost curve has
already begun to increase. Given how short this interval is likely to be, the chances are remote. Even if it
occurs, the welfare losses are likely to be small. Over that range, the average cost curve would
necessarily be nearly horizontal, and at the point of tangency the demand curve will be nearly horizontal
as well. When this occurs, the spread between marginal cost and average cost is likely to be quite narrow,
and the deadweight loss is likely to be quite small.
67
See supra note 23 and accompanying text.
23
approach is the fact that imperfect price discrimination could also
theoretically cause deadweight loss either to increase or decrease.
Consequently, the public goods approach holds that the ultimate welfare
implications of imperfect price discrimination are ambiguous.
The analysis under monopolistic competition thus places far more
importance of the appropriation of surplus from all sources without regard to
whether it comes form incremental sales of the transfer of wealth from
consumers to producers. Far from being economically irrelevant, as
suggested by conventional economic theory, under monopolistic competition
the transfer of surplus from consumers to producers is a necessary condition
for dynamic efficiency.
In addition, it lacks even distributional
consequences in the long run, since free entry will dissipate any profits that
initially accrue to the producer. Any rents initially captured by producers
will eventually accrue to the benefit of consumers in the form of increased
product diversity.
b. The Indirect Relationship Between Appropriability and Static
Efficiency
Product differentiation theory thus casts a new light on the role that
appropriability plays in fostering dynamic efficiency. What to date has been
completely ignored in the literature is an appreciation of the role that
appropriability can play in promoting static efficiency even when it does not
cause output to increase.
As noted earlier, increasing total output represented the only means
identified by the public goods approach for promoting static efficiency.
Product differentiation theory, in contrast, allows for a new and more
dynamic way to promote static efficiency through new product entry. Recall
that entry by new products causes the demand curve facing each of the
incumbent firms to shift backwards and to flatten. Because monopolistically
competitive markets reach equilibrium where the demand curve is tangent to
the average cost curve, this necessarily implies that entry will necessarily
cause the market to reach a new equilibrium at a point where the spread
between price and marginal cost is narrower. This has the inevitable effect
of causing deadweight loss to decrease. As entry drives demand towards
perfect elasticity, it will also drive deadweight loss towards zero.
Those unaccustomed to viewing markets through the lens of product
differentiation may find some aspects of this new equilibrium somewhat
puzzling. For example, the fact that demand is shifting backwards
necessarily implies that the output and total surplus generated by any
particular work will fall. The key to unlocking this puzzle is keeping in mind
that modeling market interactions at the firm rather than the industry level
makes total welfare a function of the number of works created as well as the
surplus generated by any particular work. So long as firms can appropriate a
24
sufficient proportion of the available surplus, the welfare gains from new
entry should more than compensate for the reductions in surplus generated
by the incumbents.
Stimulating entry thus represents an independent way to promote static
efficiency that only becomes apparent when the possibility of product
differentiation is acknowledged. Unfortunately, it is also a basis for
improving welfare that the public goods approach is poorly situated to take
into account. When products are assumed to be homogeneous, entry is
unlikely to yield no welfare benefits.
Indeed, when products are
homogeneous and nonrival, any attempt at duplicative entry simply wastes
resources. Moreover, the tendency towards natural monopoly created by
declining-cost structure strongly suggests that no such entry would even be
viable.
The linchpin to promoting static efficiency by inducing entry is
appropriability. Initially, increases in appropriability will cause the short-run
profits of the incumbent producers to increase. As a result, strengthening the
supracompetitive returns of the existing players may seem to be a somewhat
counterintuitive way to redress what amounts to a problem of excessive
concentration. But it is precisely those short-run profits that provide the
signals to potential entrants that entry is feasible and that spurs them to incur
the fixed cost investments that must be undertaken before a new product can
be offered. Moreover, so long as entry remains free, new works will
continue to appear until all supracompetitive returns dissipate. The long-run
effect of enhancing appropriability is thus greater competition and diversity
rather than sustainable profits. Indeed, failure to enhance appropriability will
only serve to deter entry by reducing the number of firms that can enter
successfully. As a result, any undue limitation on appropriability can have
the perverse effect of cementing an excessively concentrated market
structure into place.
Product differentiation theory thus reveals that static and dynamic
efficiency are linked by a richer and more complex set of interactions than
the simple conflict between “access” and “incentives” posited by the public
goods approach would lead one to believe. It also underscores that transfers
of wealth from consumers to producers can play as key a role in promoting
static efficiency as it does in promoting dynamic efficiency.
c. Demand Diversion as a Countervailing Consideration
The analysis advanced thus far would appear to suggest that steps that
tend to increase author’s ability to appropriate surplus tend to enhance
welfare. Indeed, were appropriability the only relevant consideration,
copyright policy would devolve into a simple matter of allowing authors to
capture as much revenue as possible. This simple result is complicated by
the existence of another dynamic efficiency consideration inherent in product
25
differentiation that tends to militate in the other direction: the distinction
between “demand creation” and “demand diversion.”68
As stated earlier, the basic welfare analysis of dynamic efficiency dictates
that a product be created whenever the benefits created by the product
exceed the costs required to produce it. When the actor in question is a
monopolist selling homogeneous products as suggested by the public goods
approach, all increases in output necessarily result from “demand creation,”
i.e., sales to new customers who were not previously purchasing any
products. A simple profitability constraint should effectively ensure that
only welfare-enhancing works are created. That is because any revenue
captured by a new entrant would necessarily represent an incremental
increase in total surplus. The fact that a firm was able to breakeven would
by itself be sufficient evidence that the surplus created by the work exceeded
the fixed costs needed to produce it.
A different situation obtains when differentiated products are involved.
Because products act as imperfect substitutes for one another, it is quite
possible that not all of the sales garnered by a new entrant necessarily come
from demand creation. It remains possible that some sales result from
“demand diversion,” i.e., the transfer of surplus from an incumbent producer
to the new entrant. When demand diversion is present, imposing a
breakeven constraint is no longer sufficient to ensure that authors will only
produce new works when doing so will enhance welfare. Under the welfare
criterion described above, an author should produce a new work only when
the surplus attributable to demand creation (i.e., surplus that represents an
increase in total welfare) is larger than the fixed costs that must be incurred
for the work to be produced. The problem is that a profit-maximizing author
will produce whenever the total surplus it captures exceeds fixed cost,
without regard to whether that surplus came from demand creation or
demand diversion.
As a result, it is quite possible that the presence of demand diversion
might lead an author to produce a work even though the incremental surplus
created by the work fell short of the fixed costs needed to produce it. Such
an author would simply finance the fixed costs that were not covered by
incremental surplus with surplus cannibalized from other producers already
in the market. In the words of Gregory Mankiw and Michael Whinston,
demand diversion “drives a wedge between the entrant’s evaluation of the
desirability of his entry “and that of an imaginary, omniscient and
omnipotent social planner.69 Under these circumstances, am author may well
68
The terminology used in this discussion is taken from Borenstein, supra note 50, at 388-89. For
similar analyses using other terminology, see Koenker & Perry, supra note 61, at 226-27; N. Gregory
Mankiw & Michael D. Whinston, Free Entry and Social Inefficiency, 17 RAND J. ECON. 48, 49, 52, 54-55
(1986); Spence, supra note 61, at 230-31.
69
Mankiw & Whinston, supra note 68, at 49.
26
find it profitable to create a new work even though the incurrence of the
fixed costs needed to create the work would constitute a waste of resources.
The use of a Hotelling-style spatial model provides an even more intuitive
illustration of these effects. In Hotelling’s original model, stores choose
locations along a geographic space, such as a main street of a city, along
which customers were uniformly distributed.70 Because of transportation
costs, customers derive greater utility from shopping at stores that are closer.
Utility declines with distance from the store until completely consumed by
transportation costs, at which point the customer decides not to shop.
Hotelling recognized that the same framework could be used to model a
group of products distributed along characteristics space.71 For example,
instead of deciding where to locate along a geographic continuum,
manufacturers of a product like apple cider could face a decision of where to
produce along a spectrum running from sweet to sour. Customers would
decide to purchase based on the proximity of a particular product to their
ideal flavor of cider, with the decline in utility as the good consumed
diverges from the customers’ most preferred bundle of characteristics taking
the place of transportation costs.
Examples of two such product spaces are depicted in Figure 4. The
horizontal dimension depicts where along the sweet-sour continuum that a
particular firm has chosen to locate itself. The vertical dimension in the
graph representing the net surplus resulting from purchases by individuals
whose ideal preferences are located at that particular location. Total surplus
is thus represented by the area of each triangle. The two exterior triangles
represent the incumbent firms; the interior triangle represents the new
entrant.
The left-hand graph depicts products that are relatively poor substitutes
for one another, in that customers’ utility falls off fairly quickly as the
distance from their ideal product characteristics. In that group, all of the
surplus captured by the entering firm represents demand creation (i.e., new
sales). Consequently, the surplus it captures represents an accurate reflection
of the social benefits created by its product. Any such product able to cover
its own costs will necessarily be welfare enhancing.
70
71
Hotelling, supra note 43, at _.
Id. at _; see also Kaldor, supra note 42, at 37.
27
Figure 4
Spatial Representation of Entry With and Without Demand Diversion
entrant
sweet
entrant
sweet
sour
sour
The right-hand graph portrays products that are relatively good substitutes
for one another, with the adjacent firms splitting the areas that overlap. In
this case, part of the surplus captured by the new entrant comes from demand
creation (represented by the darkly shaded, diamond-shaped area in the
middle of the graph), which again represents the new product’s incremental
contribution to total welfare. Part of the surplus captured by the new entrant
also consists of demand diversion (represented by the two small, triangular,
cross-hatched areas). If the fixed costs of entry lie somewhere between the
surplus resulting from demand creation and the total surplus captured by the
entrant (including both demand creation and demand diversion), the firm will
enter even though doing so will cause total surplus to fall. Under these
circumstances, the profitability constraint is not sufficient to guarantee that
only welfare enhancing works are created. On the contrary, it is quite
possible that equilibrium may yield too much diversity.
d. Striking the Optimal Balance
Whether the equilibrium will maximize dynamic efficiency thus depends
on the balance of two countervailing factors. On the one hand, the inevitable
slippage in the appropriability of the total surplus created by any work will
28
tend to lead to the production of too few new works. On the other hand, the
possibility of demand diversion tends to stimulate entry even when doing so
wastes resources, which in turn tends to lead to the production of too many
new works. The net effect on dynamic efficiency depends upon which of
these two forces dominates the other. Ultimately, this is an empirical
question, and a rather complex one at that.
Theory does provide some guidance as to the relevant considerations.72
The factors tending to reduce appropriability have been discussed earlier73
and include high fixed costs and steep inverse demand functions (which is a
concept somewhat related with low own-price elasticity of demand). The
dangers of demand diversion are the highest when cross-price elasticities of
demand are high. Although the latter two concepts are conceptually distinct,
they can be loosely subsumed in the concept of substitutability.
These insights could be integrated into copyright through the use of a
sliding scale. Simply put, markets for differentiated products tend to produce
too many products when substitutability is high and fixed costs are low. It is
thus arguable that works exhibit such qualities should receive a lesser degree
of copyright protection. Conversely, markets for differentiated products are
likely to produce too few products when substitutability is low and fixed
costs are high. It is when these criteria are met that copyright protection
should be at its broadest.
III. THE DIFFERENTIATED PRODUCTS APPROACH TO COPYRIGHT:
ASYMMETRIC PREFERENCE MODELS
Product differentiation theory would thus appear to offer a useful
framework for fundamentally reconceptualizing copyright. It helps explain
many of the descriptive features of markets for creative works that seem
paradoxical when viewed through the lens of public goods economics. It
also opens up the policy space by raising the possibility of first-best
outcomes that can reconcile the supposedly irreconcilable conflict between
static and dynamic efficiency. It illustrates the role that short-run profits and
entry play in mediating between static and dynamic efficiency, in the process
turning the transfer of surplus from consumers to producers—a factor
generally regarded to be immaterial to welfare—into a necessary condition
for efficiency. Lastly, it renders the distributional concerns that have driven
many of the debates around copyright largely irrelevant by providing strong
reasons to believe that profits will not be sustainable over the long run in
most cases. Indeed, the theory indicates that any such profits that exist in the
short run will accrue to the benefit of consumers in the form of increased
product diversity.
72
See Koenker & Perry, supra note 61, at 226-27; Spence, supra note 61, at 230-31; Spence, supra
note 46, at 410, 413.
73
See supra notes 63-64 and accompanying text.
29
The overall thrust of these results would appear to be profound. Product
differentiation theory does sound one note of caution, however. Many of the
key aspects of the equilibrium under monopolistic competition depend on the
assumption that consumers’ preferences are spread symmetrically across all
of the available products. Understanding product differentiation theory’s
implications for copyright thus require an appreciation of the consequences
of relaxing this assumption.
A. Relaxing the Symmetric Preferences Assumption
Understanding the role that asymmetric preferences plays begins an
understanding the relationship between fixed costs and entry. Consider a
monopolistically competitive industry exhibiting the symmetric preferences
originally posited by Chamberlin. Because consumer preferences are spread
equally across all of these products each of them, each producer captures the
same amount of the surplus, represented by the total size of the market
divided by the number of firms participating in the market. Suppose now
that a new firm enters the market. Entry causes the amount of surplus
captured by each firm to become smaller, as the denominator in the abovedescribed fraction. Additional entry will continue until the amount of
surplus captured by each firm no longer exceeds the fixed costs needed to
support entry.
The number of firms in the market thus depends upon the size of the fixed
costs relative to the overall market. In fact, the number of firms can be
calculated directly by dividing the total surplus available by the size of the
fixed costs. The larger the fixed costs, the smaller the number of firms
operating in the industry. The number of firms will not be overly restrictive
so long as the size the overall market is large. As noted earlier,74 the relative
size of the fixed costs also determines the significance of the sustainable
profits associated with the so-called “integer problem.” Preservation of free
entry and the absence of profits thus depend upon the fact that the relevant
market is large in comparison to fixed cost. The symmetry assumption
ensures that the numerator of the market-to-fixed cost ratio is relatively
large.
The scenario changes, however, when one relaxes the symmetry
assumption and allows for the possibility that products will not exhibit the
same degree of substitutability with respect to all products.75 Variations in
74
See supra note 54 and accompanying text.
This discussion follows the seminal analysis offered by Kaldor, supra note 42, at 37-40. It bears
emphasizing that symmetry of consumer preferences across products (i.e., uniformity of cross-price
elasticities) is only one of several symmetry assumptions entertained by Chamberlin. His model also
assumed that every firm confronted the same demand curve as well as the same cost structure. Relaxation
of these assumptions do not raise any significant problems. Variations in demand curves would simply
cause tangency to occur at different prices for different firms without preventing each firm from reaching
equilibrium. A similar argument neutralizes the effect of variations in cost structure. Id. at 43-45. It is
75
30
cross-price elasticity create the possibility that firms will face greater
competition from works which are relatively good substitutes and less
competition from those which are relatively poor substitutes. If so, it can no
longer be assumed that the impact of entry by a new firm will be spread
evenly across all of the incumbents and rendered negligible with respect to
any particular firm. Instead, it suggests that the degree of congestion (and
thus the freedom of entry and sustainability of profits) will vary across the
overall market.
Indeed, if consumers used a single parameter to distinguish among
products, it would be quite natural to depict these asymmetric preferences
spatially by assigning each of the competing products a location along a
particular continuum. Products would compete more fiercely with their
nearer neighbors and less fiercely with those farther off. If the degree of
differentiation among products is particularly high, gaps in the product
continuum may divide the market into subsegments. Division of the market
into subsegments accentuates the role played by fixed costs. This is because
the relevant numerator in the market-to-fixed cost proportion is now the size
of the subsegment rather than the size of the overall market.
The possibility of market segmentation thus raises the possibility that the
integer problem may arise with respect to multiple subsegments of the
overall market. As a result, allowing for the possibility that consumer
preferences may be asymmetric across products may reduce the number of
firms that can operate in the market by heightening the effect of fixed cost
indivisibilities. It also creates the real possibility of more widescale
sustainable profits, as the single “large economy” is chopped into a larger
number of “small economies.” Interestingly, the market need not be divided
into discrete subsegments in order for this effect to occur. Variations in the
density of firms across the product space can balkanize the industry into a
chain of “overlapping oligopolies,” each comprised of a small number of
firms engaged in localized competition regardless of how many firms are
operating in the overall market. This can give rise to the same problems
even in the absence of actual gaps in the product continuum.76
The magnitude of this problem is thus not simply determined by the size
of the fixed costs. It also depends on the cross-price elasticity of demand
between products. The lower the degree of substitutability, the more
segmented the market is likely to be. It bears mentioning that models that
limit product differentiation to a single dimension are particularly susceptible
to the existence of sustainable profits that will not be dissipated by new
only the assumption with respect to consumer preferences and the manner in which that assumption
spreads the effects of entry evenly across all market participants that significantly alters the analysis.
76
This provides an answer to the criticism that Chamberlinian product groups are nothing more
than Marshallian industries. See FRIEDMAN, supra note 26, at 38-39; STIGLER, supra note 26, at 17. It
suggests that products that do not compete directly with one another may nonetheless be linked together
through a chain of competitive products.
31
entry. This is because such models limit each firm to competing only with
their immediately adjacent neighbors. A very different result obtains if
product differentiation occurs along multiple dimensions. For example, if
competition takes place on three dimensions, each product may compete with
as many as six other products. If competition expands to four dimensions,
firms may compete on average with as many as half of all of the firms
operating in the industry.77 Multidimensionality of competition can thus
largely eliminate the localized nature of differentiated products competition
largely disappears.
B. Implications
The asymmetric preferences branch of product differentiation theory
paints a far less rosy picture than does the symmetric preferences branch.
Allowing consumer preferences to vary across products has the effect of
reducing the size of the relevant market for any particular work to something
less than the industry as a whole. This in turn increases the likelihood that
fixed costs will constitute a barrier to entry and enhances the possibility of
sustainable profits.
If these effects arise with respect to multiple
subsegments, these profits may be quite substantial.78
Allowing for the possibility of asymmetric preferences only serves to
heighten the importance of the two factors that the symmetric preferences
models already identified as crucial: the degree of substitutability and the
magnitude of fixed costs. It also intensifies the need to “thicken” the market
in those subsegments where substitutability is relatively low and fixed costs
are relatively high. The interrelationship between static and dynamic
efficiency identified above indicates that the best way to thicken markets is
by increasing the amount of the surplus appropriable by authors. That this
should be the proper response to the presence of barriers to entry and
supracompetitive profits may at first glance seem somewhat incongruous. It
seems less so once one recognizes the manner in which entry by new
products promotes both static and dynamic efficiency when products are
differentiated as well as the critical role that the appropriability of surplus
plays in stimulating that entry.
77
G.C. Archibald & G. Rosenbluth, The “New” Theory of Consumer Demand and Monopolistic
Competition, 80 Q.J. ECON. 569 (1975).
78
See B.C. Eaton & R.G. Lipsey, The Non-Uniqueness of Equilibrium in the Loschian Location
Model, 66 AM. ECON. REV. 77 (1976).
32
IV.APPLICATION OF PRODUCT DIFFERENTIATION THEORY TO SPECIFIC
COPYRIGHT ISSUES
A. Facilitation of Price Discrimination
Perhaps the most salient issue in current copyright scholarship focus on
the extent to which copyright and related doctrines should be structured to
facilitate price discrimination.79 As noted earlier, the public goods approach
focuses almost exclusively on whether price discrimination will mitigate the
welfare reductions associated with deadweight loss. In the process, it adopts
the conclusion drawn by traditional monopoly theory to assert that imperfect
price discrimination may cause output to increase or decrease, depending on
the own-price elasticities of demand of the various market segments. The
conventional wisdom is thus that the welfare impact of imperfect price
discrimination is ambiguous.
The analysis advanced in this article underscores the crucial role played
the appropriability of surplus plays in maximizing total welfare. If the
relevant market involved homogeneous products, the answer would be
relatively straightforward: any degree of nonappropriability would prevent
some welfare-enhancing works from being produced. The conclusions differ
quite starkly when the products involved are differentiated products. The
possibility of demand diversion raises the possibility that complete
appropriability would lead to excess diversity.
Product differentiation theory thus suggests that the strength of price
discrimination should be governed by a sliding scale that facilitates a greater
degree of price discrimination when the market is most likely to reach
equilibrium at a point where the number of works created is inefficiently
low. It is under these circumstances that the enhanced ability to appropriate
surplus made possible by price discrimination is the most justified. The
market tends to produce an inefficiently low number of goods when fixed
costs are relatively high and relatively few substitutes for the product exist.
Unfortunately, these are also precisely the circumstances under which entry
is likely to be the most difficult and when the incumbents have the greatest
chance of earning sustainable profits.
It may seem counterintuitive to promote price discrimination in those
instances when markets are most likely to be concentrated and most likely to
generate supracompetitive returns. From the narrow standpoint of static
efficiency, enhancing the market power of authors in such situations would
appear to be anomalous. This apparent incongruity disappears when the
perspective is augmented to include dynamic efficiency considerations. The
concentration in the market and the presence of economic profits is a signal
that that subsegment of the overall market is too horizontally concentrated.
79
See supra note 7 and accompanying text.
33
The logical and most administratively sustainable solution to this problem is
to encourage entry into this segment by new competitors. The best way to
encourage such entry is by increasing appropriability. Short-run profits may
increase, but it is precisely those profits that signals potential entrants that
entry is needed and that spurs them to undertake the fixed cost investments
needed to do so. Conversely, limiting producers’ ability to appropriate
surplus would make it harder for new firms to enter by heightening the extent
to which fixed costs would act as barriers to entry. Under such
circumstances, limiting price discrimination would have the perverse effect
of locking the existing, inefficient, excessively concentrated market structure
into place.
The logic of entry under monopolistic competition also suggests ways in
which entry by new competitors has the indirect of lessening the static
inefficiency caused by supramarginal cost pricing. As noted earlier,80
competitive entry causes the demand curve facing incumbent firms to shift
backwards and to become more elastic. This flattening of the demand curve
will necessarily reduce deadweight loss by causing the spread between price
and marginal cost to narrow.
This dynamic underscores the role that entry by new products plays in
mediating between static and dynamic efficiency. Even in the face of
asymmetric consumer preferences, entry by new firms should cause each
firm’s demand curve should flatten as more substitutes become available.
This will have the simultaneous effect of increasing appropriability, reducing
deadweight loss, and dissipating supracompetitive profits. This complex
relationship between static and dynamic efficiency is largely obscured when
copyright are viewed through the lens of public good economics and
conventional monopoly pricing, which have the inexorable tendency of
placing static and dynamic efficiency in tension with one another. Shifting
to a differentiated products perspective brings what otherwise appear to be
conflicting considerations into alignment.
B. The Fair Use Doctrine
The mainstream economic justification for the fair use doctrine regards it
as a means for compensating for market failure caused by transaction costs.81
In the tradition of Calabresi and Melamed,82 this perspective posits that the
government should mandate access to a work when friction in the bargaining
process prevents a welfare-enhancing transaction from occurring. The
current debate regarding fair use centers on whether the reduction in
80
See supra pp. 15-16
See Wendy J. Gordon, Fair Use as Market Failure: A Structural and Economic Analysis of the
Betamax Case and Its Predecessors, 82 COLUM. L. REV. 1600, 1628-30 (1982).
82
Guido Calabresi & A. Douglas Melamed, Property Rules, Liability Rules and Inalienability:
One View of the Cathedral, 85 HARV. L. REV. 1089 (1972).
81
34
transaction costs made possible by improvements in technology justifies
narrowing the doctrine’s scope.83
Product differentiation theory casts a new spin on the economic approach
to the fair use doctrine. Fair use seems well positioned to promote static
efficiency by making it easier to satisfy low-value uses. At the same time, it
tends to harm dynamic efficiency by reducing the author’s ability to
appropriate surplus. As noted earlier, however, there are circumstances
under which reductions in appropriability actually enhance welfare, i.e.,
when fixed costs are low and substitutes are readily available. When those
circumstances exist, allowing a greater degree of fair use is less likely to
harm total welfare.
Application of product differentiation theory would thus transform fair
use into a sliding scale keyed to fixed costs and the availability of substitutes.
Fair use would be broadest when the relevant elasticities and fixed costs
make substitutes the most readily available. This differs from the transaction
cost regime in that technological change no longer becomes a one-way
ratchet that inexorably tends to contract the scope of fair use. Rather, it ties
fair use to a more structural type of market failure. This is not to say that
those focusing on the economics of differentiated products would find
frictional concerns to be irrelevant. On the contrary, such considerations
would necessarily play an important role in the design of any remedy. As a
result, product differentiation and the transaction cost theory in some ways
serve as complementary rather than alternative approaches. That said, it does
appear that the focus of product differentiation theory provides certain
insights that transaction cost approach is not well designed to recognize.
This analysis of fair use replicates the irony that discussed in the context
of price discrimination that copyright protection would appear to be the
strongest under those circumstances in which, at least from the standpoint of
static efficiency, it is needed the least. As noted earlier, this apparent
contradiction simply reflects the complex interaction between static and
dynamic efficiency.
C. The Scope of Copyright Protection
[Forthcoming]
83
See, e.g., PAUL GOLDSTEIN, COPYRIGHT’S HIGHWAY 224; Tom W. Bell, Fair Use vs. Fared
Use: The Impact of Automated Rights Management on Copyright’s Fair Use Doctrine, 76 N.C. L. REV.
557 (1998); Trotter Hardy, Property (and Copyright) in Cyberspace, 1996 U. CHI. LEGAL FORUM. 215;
Edmund W. Kitch, Can the Internet Shrink Fair Use?, 78 NEB. L. REV. 880 (1999); Robert P. Merges,
The End of Friction? Property Rights and Contract in the “Newtonian” World of On-Line Commerce, 12
BERKELEY TECH. L.J. 115 (1997).
35
CONCLUSION
In the final analysis, product differentiation appears offer significant
promise as a new way to reconceptualize copyright law. Not only does it
offer an explanation for a number of features that the conventional, public
goods approach cannot. Allowing for competition in more than one
dimension opens up the policy space in ways that make first-best solutions
possible. Product differentiation theory also provides a way to formalize the
consideration of both static and dynamic efficiency and yields insights into
their deep structural interrelationship. What results is a sliding-scale
approach in which the breadth of the various copyright doctrines depend
upon contextual factors. There is some irony in the fact that copyright tends
to be the broadest when high fixed costs and the low degree of
substitutability causes the market to become the most restricted, but the
paradox disappears once one has a better understanding of the complex
interaction between static and dynamic efficiency.
I confess that I have approached this project as a purely theoretical
exercise. Consequently, I have paid relatively little attention to the
implementability of the policy recommendations that result from it. I thus
must acknowledge that my analysis is vulnerable to a whole host of
complaints about administrability, many of which have been raised before.84
By its nature, product differentiation theory is an extremely empirical and
contextual enterprise that does not lend itself to simple policy inferences.
Chamberlinian product groups are notoriously hard to define, and the slidingscale approach that I have suggested will be even harder to implement.
Even the critics of product differentiation theory acknowledge that the
proper measure of a theory is not its tractability, but rather its ability to offer
predictions that make a better fit with reality.85 For the reasons laid out in
Part I.B., I think that that criterion is met.86 As a result, I do not necessarily
regard the fact that the theory requires a nuanced assessment of the empirical
context as representing a basis for rejecting the theory.87
84
See FRIEDMAN, supra note 26, at 38-39; STIGLER, supra note 26. For a useful overview of this
debate, see Archibald, supra note 26, at 2-8.
85
See FRIEDMAN, supra note 26, at 8-9; STIGLER, supra note 26, at 23; Archibald, supra note 26, at
2-5.
86
See also Klein, supra note 50, at 75 n.59.
87
See EDWARD HASTINGS CHAMBERLIN, TOWARDS A MORE GENERAL THEORY OF VALUE 305
(1957); Archibald, supra note 26, at 2 n.3.
36